Features
Role-based Agent Design
Define agents with specific roles, goals, and backstories to simulate a team with diverse expertise.
<Team>
<Agent role="designer">
<Task>design awesome landing page</Task>
</Agent>
<Agent role="developer">
<Task>develop awesome landing page</Task>
</Agent>
</Team>
Autonomous Inter-Agent Delegation
Agents can delegate tasks and collaborate autonomously, improving collective problem-solving capabilities
Flexible Task Management
Assign tasks to agents with customizable tools, allowing for dynamic task allocation and execution
Multiple Types of Processes
The current implementation supports sequential task execution, with plans to introduce more complex processes like consensus-based and hierarchical task management.
<Team process="sequential">
// ...
</Team>
React-compatible Framework
Built in JSX, enabling integration with the React ecosystem and leveraging its features for UI and state management.
/** @jsxImportSource ai-jsx/react */
import * as AI from "ai-jsx/react";
const ReactClientComponent = () => {
return (
<AI.JSX>
<Team process="sequential">
// ...
</Team>
</AI.JSX>
)
}
JSX Syntax for AI Constructs
Utilize familiar JSX syntax to compose AI agents and tasks, making it intuitive for every developer. See more
Integration with AI.JSX
Works in conjunction with AI.JSX, a framework for building AI applications, to provide a seamless development experience for AI-driven projects. See more
Extensible Tooling System
Agents can be equipped with tools that enhance their capabilities, allowing for extensible and modular functionality.
Model Selection
Agents can be configured to use different AI models and providers, offering flexibility in the AI capabilities.
<Team>
<Agent
role="designer"
model="gpt-3.5-turbo" // <---- Pick any model
provider="openai" // <---- Pick any provider
>
<Task>design awesome landing page</Task>
</Agent>
<Agent
role="developer"
model="mistral"
provider="ollama"
>
<Task>develop awesome landing page</Task>
</Agent>
</Team>